fix: fix Rust v1.78.0 clippy warnings

[Boshen]

No description provided.

[Boshen]

warning: this bound is already specified as the supertrait of SourceCode

[Boshen]

🤔 Rust v1.78.0 also surfaced 2 unspecified features

error: unexpected `cfg` condition name: `doc_cfg`
 --> src/eyreish/mod.rs:1:13
  |
1 | #![cfg_attr(doc_cfg, feature(doc_cfg))]
  |             ^^^^^^^
  |

error: unexpected `cfg` condition name: `track_caller`
  --> src/eyreish/mod.rs:87:12
   |
87 | #[cfg_attr(track_caller, track_caller)]
   |            ^^^^^^^^^^^^
   |

[Benjamin-L]

I think the #[cfg_attr(track_caller, ...)] usage was inherited from eyre. They attempt to probe stable track_caller support in the build script, and then set cfg=track_caller if it's available. We aren't doing this, but are using track_caller unconditionally already, and track_caller is available in our MSRV. I looked through all the uses of #[cfg(track_caller)] in miette, and believe we can just enable them all by default without a breaking change. I'd be in favor of doing that, although we should test it carefully.

doc_cfg is in a similar situation, where the code was inherited from eyre, and they set the cfg in the cargo manifest. We don't seem to be using #[cfg(doc)] anywhere, and this feature is also stable in our MSRV. We should just remove the :#![cfg_attr(doc_cfg, feature(doc_cfg))] line.

[Boshen]

I've added #![allow(unexpected_cfgs)] so we can unblock other PRs.

I'll create an issue for removing these unexpected cfgs.

[Boshen]

Ahh sorry I totally forgot about this PR, MIRI is failing due to

error: unexpected `cfg` condition name: `track_caller`
  --> tests/test_fmt.rs:83:12
   |
83 | #[cfg_attr(track_caller, ignore)]
   |            ^^^^^^^^^^^^

You may #[allow()] it as well.